Skip to main content

Table 2 Comparison of memory between the JA algorithm and our algorithm on an input with 8 million reads from a plant genome.

From: Efficient parallel and out of core algorithms for constructing large bi-directed de Bruijn graphs

   JA ALGO ParBiConstruct
P k RES.MEM(Mb) SHR.MEM(Mb) RES.MEM(Mb) efficiency SHR.MEM(Mb) efficiency
8 21 903.714 3.690 244.603 3.695 3.328 1.109
8 27 675.516 3.638 146.750 4.603 3.232 1.126
8 33 289.787 3.588 62.580 4.631 3.007 1.193
8 35 116.938 7.527 14.625 7.996 2.397 3.140
12 21 681.664 4.782 156.875 4.345 4.256 1.124
12 27 265.304 6.882 91.219 2.908 3.889 1.769
12 33 241.630 4.721 46.805 5.163 3.675 1.285
12 35 94.548 7.518 12.276 7.702 2.771 2.713
16 21 501.175 5.726 152.995 3.276 5.175 1.107
16 27 333.487 11.730 77.523 4.302 4.823 2.432
16 33 166.395 5.714 42.746 3.893 4.628 1.235
16 35 68.851 8.172 13.578 5.071 3.443 2.374
24 21 344.888 8.081 105.723 3.262 7.133 1.133
24 27 214.681 16.271 61.674 3.481 6.690 2.432
24 33 115.630 7.830 34.583 3.344 5.959 1.314
24 35 52.110 10.082 15.990 3.259 4.472 2.254
32 21 260.063 10.413 86.157 3.018 9.070 1.148
32 27 179.657 20.550 49.737 3.612 8.096 2.538
32 33 95.790 9.968 34.180 2.803 7.495 1.330
32 35 47.602 11.290 15.289 3.113 4.905 2.302
48 21 186.225 14.109 71.073 2.620 12.405 1.137
48 27 112.237 19.623 47.596 2.358 11.255 1.744
48 33 75.059 11.996 30.929 2.427 9.070 1.323
48 35 37.500 10.643 19.576 1.916 6.764 1.573
64 21 150.324 17.311 63.449 2.369 14.640 1.182
64 27 105.692 23.570 43.590 2.425 12.975 1.817
64 33 66.095 14.168 31.853 2.075 10.653 1.330
64 35 38.823 12.159 22.275 1.743 8.096 1.502
  1. Input Reads: 8388608 short reads, each of size 36 base pairs
  2. Average Efficiency of RESident memory across all processors = 3.622
  3. Average Efficiency of SHaRed memory across all processors 1.667